Taxonomic history

Banksia subg. Isostylis was first published by Robert Brown in his 1810 Prodromus Florae Novae Hollandiae et Insulae Van Diemen; thus its full name, with author citation, is Banksia subg. Isostylis R.Br.^[3] Brown's arrangement was the first infrageneric arrangement of Banksia, making B. subg. Isostylis Banksia's first infrageneric taxon.^[4] Brown erected B. subg. Isostylis to contain B. ilicifolia, which was then the only known Banksia with a dome-shaped inflorescence. He did not explicitly name a type species for the subgenus,^[4] but B. ilicifolia is treated as the type because it was the only member when the subgenus was published.^[1]

Twenty years later, Brown issued a supplement to his Prodromus entitled Supplementum Primum Prodromi Florae Novae Hollandiae; another nine Banksia species were published, but there was no change to the 1810 arrangement, and no new Isostylis species.^[5]

In 1846, Édouard Spach promoted B. subg. Isostylis to genus rank in his Histoire Naturelle des Vegetaux: Phanerogames. This was not accepted, and Isostylis (R.Br.) Spach is now considered a nomenclatural synonym of B. subg. Isostylis.^[6]

When Carl Meissner published his arrangement of Banksia in 1856, he demoted both of Brown's subgenera to sectional rank, maintaining B. sect. Isostylis (R.Br.) Meisn. as a monospecific taxon containing only B. ilicifolia. Meissner's rank and circumscription of Isostylis was retained by George Bentham in his 1870 arrangement for Flora Australiensis, but Bentham also published a putative variety of B. ilicifolia, B. i. var. integrifolia, based on specimens collected by Ludwig Preiss near the Swan River in Western Australia. This was later overturned.

In 1905, James Britten challenged the genus name Banksia, on the grounds that Banksia J.R.Forst & G.Forst had precedent over Banksia L.f. Britten adopted the name Isostylis for the entire genus, republishing the names Isostylis dentata (L.f.) Britten for B. dentata (Tropical Banksia), Isostylis ericifolia (L.f.) Britten for B. ericifolia (Heath-leaved Banksia), Isostylis integrifolia (L.f.) Britten for B. integrifolia (Coast Banksia) and Isostylis serrata (L.f.) Britten for B. serrata (Saw Banksia). This challenge failed, Banksia L.f. was eventually conserved, and his four names are now considered taxonomic synonyms of their respective names under Banksia L.f. None of them are considered members of Isostylis.

The next change to Isostylis came in 1981, when Alex George promoted it back to subgenus rank, and published a second species, B. cuneata. In discussing the subgenus, George commented that there had been calls to transfer Isostylis into Dryandra, which was then a distinct genus. He argued, however, that the similarities between Isostylis and Dryandra were largely superficial, whereas the similarities with Banksia were much more important taxonomically. His conclusion was that the taxon should remain in Banksia, although he did not rule out promoting it into a separate genus. A third B. subg. Isostylis species, B. oligantha, was published by George in 1988.

In 1996, Kevin Thiele and Pauline Ladiges published a revised arrangement based on a cladistic analysis of morphological characters of Banksia. They took up the question of an affinity of Isostylis with Dryandra, finding George's arguments unconvincing but failing to find any further evidence for or against Isostylis's placement within Banksia. They eventually accepted both of George's subgenera, using each as an outgroup in the analysis of the other. Thus their analysis yielded no information about the circumscription and placement of Isostylis, and their arrangement maintained Isostylis as a subgenus.^[7]

Thiele and Ladiges' arrangement was not accepted by George, and was largely discarded by him in his 1999 arrangement. The placement and circumscription of B. subg. Isostylis was unaffected and can be summarised as follows:^[2]

Banksia

B. subg. Banksia (3 sections, 11 series, 73 species, 11 subspecies, 14 varieties)

B. subg. Isostylis

B. ilicifolia

B. oligantha

B. cuneata

Since 1998, Austin Mast has been publishing results of ongoing cladistic analyses of DNA sequence data for the subtribe Banksiinae. His analyses has provided compelling evidence for the paraphyly of Banksia with respect to Dryandra, and suggest an overall phylogeny that is very greatly different from George's arrangement. Mast's results clearly recognise Isostylis as a distinct clade, but place it in a fairly recent position, within a clade that also contains B. elegans (Elegant Banksia) and B. attenuata (Candlestick Banksia). It falls a substantial distance from Dryandra, suggesting that similarities between those two groups are indeed superficial.^[8][9][10]

Early in 2007, Mast and Thiele initiated a rearrangement of Banksia by merging Dryandra into it, and publishing B. subg. Spathulatae for the species having spoon-shaped cotyledons. They foreshadowed publishing a full arrangement once DNA sampling of Dryandra was complete; in the meantime, B. subg. Isostylis has been set aside. If maintained at all in Mast and Thiele's forthcoming arrangement, it will be at a lesser rank than subgenus.^[11]

Relationships within B. subg. Isostylis still remain unclear. Though Mast's studies found B. cuneata to be the most basal of the three species,^[9] a 2004 study of genetic divergence within the subgenus yielded both other possibilities: some analyses suggested B. ilicifolia as basal, while others suggested B. oligantha. Further complicating the situation is the existence of a population of B. cuneata having both genetic and phenetic affinities with B. oligantha. The origin of this population is unknown. It might have arisen through hybridisation, or it may be a transitional or even ancestral form. Finally, biogeographical factors suggest that B. ilicifolia would be the most basal of the three species: it occurs in the High Rainfall Zone where relictual species are most common, whereas the others are restricted to the Transitional Rainfall Zone, where more recently evolved species are most common.^[12]